It is well known that HID lamps commonly operate in three modes, i.e., a cold or starting mode, an operating or steady mode and a hot restrike mode. During the cold or starting mode, a relatively high value of ac or dc type starting or ignitor voltage (e.g., 25KV) is applied across the lamp electrodes to first place the gasses of the lamp into a suitable ionized condition for striking or initiating a glow breakdown state. This glow breakdown state which consists of an early glow, Townsend glow and anamolous glow, is followed by a delay time period to allow a transition into an arc condition between the electrodes of the lamp during the glow-to-arc state.
The initial or early glow state is characterized by a very low current density and relatively high average energy per particle (i.e., E/N, wherein E is electric field and N is the number density of gas atoms in the discharge). Generally the total pressure is fairly low being determined by the rare gas R.sub.g starting gas and the partial pressure of mercury is quite low, being determined by the ambient temperature. Since the excitation cross-section for e.sup.- +R.sub.g is smaller than for e.sup.- +Hg in the early stages of glow, the electrons undergo elastic collisions with R.sub.g and gain considerable energy until they either excite R.sub.g or Hg. The Hg excitation is more likely to lead to ionization which produces an additional electron-ion pair. The majority of the collisions impart low energy transfer or gas heating.
As the early glow state proceeds to the Townsend glow state, the current increases but the current density remains low. During this state, the supply of electrons originate from surface field emission which occurs with a relatively high field strength, i.e., greater than 20 volts. This period is particularly damaging to the tungsten electrodes since any ionized gas atoms are driven back to the cathode with large energy which promotes sputtering of electrode material. Thus, it is important to establish a thermionic or "hot-spot" mode of operation as quickly as possible.
The glow current will continue to increase if a sustaining voltage (not the ignitor voltage) is maintained. Usually, this initial sustaining voltage is hundreds of volts, i.e., 1000 volts/cm field strength.
Near the cathode, current continuity must take place, i.e., ions are collected and electrons emitted. However, the electrons move much more freely so that within some distance from the cathode a space charge will exist, i.e., the ions cannot get to the cathode surface as fast as the electrons can move away from the surface. This condition produces a "virtual" anode. The distribution of ions will be diffuse only because the charge density is low, i.e., the vast majority of the atoms are neutral. As an example, a fully developed discharge may only have local charge densities of 10.sup.15 /cm.sup.3 while the total number density would be 10.sup.18 /cm.sup.3. During the glow state, the density will be orders of magnitude smaller than the arc mode.
The diffuse "cloud" or sheath will shrink or move closer to the cathode as the current increases, in part due to increased ionization and due to increased repulsion within the sheath, i.e., the field gradient will increase as the sheath shrinks.
Eventually, the potential generated by the cloud is large enough to permit fairly high energy collisions of the ions with the cathode surface and upon shrinking of the sheath, the cathode becomes hot at a fairly localized point. During this mode, called the anamolous glow, the current reaches nearly the arc discharge level or greater, but the current density is still fairly small and the field strength in the vicinity of the cathode surface is high.
As the electrode surface temperature increases, the liberation of electrons becomes easier due to the reduction in the barrier potential of the electrode material. The production of electrons is still controlled by the surface area and the electrode is said to be a cold emitter.
At some point, the electrode will emit electrons by virtue of its high temperature, the so-called thermionic mode. This mode permits may times the current density so that the sheath collapses to a small spot close to the surface of the cathode. In addition, the potential drops to 10-15 volts depending on the material properties of the electrode. The thermionic mode still requires some potential gradient to remove electrons and this potential is referred to as the work function.
So during the glow, the potential across the electrodes or gap can be maintained at hundreds of volts with the majority of the drop occurring around the cathode and little field strength in the remaining region of the electrode gap. However, when the spot is formed, producing glow-to-arc, the voltage across the gap is controlled by the electron and ion mobility, which is number density or pressure dependent. If current is limited by a regulator or ballast, then the voltage across the gap will increase until the capsule body has come into thermal equilibrium.
In the operating mode, the arc discharge of the lamp generates a desired light output and a relatively low or moderate voltage occurs across the electrodes of the lamp in response to a suitable arc discharge current as established by the ballast or operating circuit related to the lamp.
The hot restrike mode occurs when the arc discharge of the lamp fails or extinguishes for some reason such as a momentary interruption of the current supplied to the lamp. If the arc discharge extinguishes resulting in a loss in light output, the lamp may be permitted to cool for a period of time before the arc condition can be restarted by the relatively high starting voltage.
Initiation of the early glow state during the hot restrike mode requires higher external potentials because the total pressure within the lamp is much higher than during the cold starting mode. This higher gas pressure inhibits the sustained glow (i.e., Townsend glow). In general, the glow "time" will be longer since it is harder to build up the necessary current density to form the thermionic mode.
During the eventual glow-to-arc mode when the pressure is high and particularly while the metal halide salt is liquid, the cathode spot mode may terminate at locations other than the tip of the electrode. The persistence of arc formation at locations such as the press interface or liquid salt is very temporary since these regions cannot deliver sustained electron flux. However, the arc spot can produce localized heating which over repeated hot start times will cause detrimental lumen maintenance.
A number of circuits have been developed in the past which specifically deal with the problem of restriking various HID lamps while they are hot so as to avoid the temporary loss of light as discussed above. Other circuits have been developed which simply wait a predetermined period of time so that restarting can be accomplished after the lamp has completely cooled.
Some of the prior art circuits are unsuitable because they simply do not work effectively, or are either relatively complex or are not reliable. More importantly, many of the these circuits are unsuitable for low voltage dc applications such as automotive headlights. In such applications, it is readily apparent that any delay in hot restarting an HID automotive headlight is intolerable.